New cannabis substitute can reduce L-dopa induced dyskinesia in Parkinson’s patients


A drug that provides the benefits obtained from medicinal cannabis without the “high” or other side effects may help to unlock a new treatment for Parkinson’s disease.

The drug – HU-308 – lessens devastating involuntary movements called dyskinesias, a side effect from years of treatment for Parkinson’s disease.

The research, published today in Neurobiology of Disease, has been conducted by the Centre for Neuroscience and Regenerative Medicine (CNRM) at the University of Technology Sydney (UTS) and the Applied Medical Research Institute of St Vincent’s Hospital Sydney.

The study shows that in mice HU-308 is as effective as amantadine, the only available treatment for dyskinesias.

Furthermore, the combination of HU-308 with amantadine is more effective than either drug used alone.

Professor Bryce Vissel, director of the CNRM and senior author of the study, said the findings present the possibility of new options for Parkinson’s patients.

“Our study suggests that a derivative of HU-308, either alone or in combination with amantadine, may be a more effective treatment for dyskinesias and a much better option than using an unproven potentially harmful substance like cannabis,” Professor Vissel said.

“Currently there is limited evidence about the effectiveness of medicinal cannabis. One problem is that no cannabis preparation is the same and cannabis has numerous effects, some of which may not be beneficial in Parkinson’s disease.”

Cannabis works on several receptors in the brain – CB1 and CB2. The psychoactive effect is caused mostly because of receptor CB1.

Professor Vissel said the HU-308 drug explored by his team works only on receptor CB2, allowing medicinal benefits to be administered without causing psychoactive effects like drowsiness or highness.

Lead author Dr Peggy Rentsch said it is unclear whether medicinal cannabis itself can help Parkinson’s patients.

“Medicinal cannabis contains different compounds, some of which make you high and which can impact a person’s normal day-to-day activities,” Dr Rentsch said.

“Our research suggests HU-308 is an important prototype drug which we believe won’t interfere with patients’ day-to-day activities.

They should maintain normal levels of mental sharpness on a treatment like this.”

The study shows that in mice HU-308 is as effective as amantadine, the only available treatment for dyskinesias.

Furthermore, the combination of HU-308 with amantadine is more effective than either drug used alone.

Professor Vissel and his team are investigating ways to block inflammation of the brain to maintain and restore memory and slow the progression for both Parkinson’s disease and Alzheimer’s disease.

“HU-308 works by reducing inflammation in the brain, affecting the neurons and immune cells.”

“In neurological disorders, the immune cells in the brain can lose supportive function with adverse stimuli – including but not limited to trauma or obesity – and become ‘activated’. Scientists at the CNRM believe that, after this activation, the immune cells backfire, kill the brain’s neurons, destroy them – and become dysfunctional.

“By reducing inflammation in the brain – such as with HU-308 – these immune cells can support normal neural function again, rather than inhibiting it.”

Study collaborator Dr Sandy Stayte said: “The fact that amantadine has its own set of side effects, may not work in the long term, and is still the only drug available on the market that is approved for dyskinesias makes our study really exciting.

“First, our study shows HU-308 is equally affective so a drug like HU-308 will be useful for those people who can’t take amantadine.

Second, for those who can tolerate amantadine, taking the combination may have even greater benefits than taking either drug alone.

That means we may end up with a much more powerful treatment than currently available by ultimately prescribing both.”

Parkinson’s disease (PD) is a neurodegenerative disorder caused by the progressive loss of dopaminergic neurons in the substantia nigra pars compacta (SNpc) and their projections into the striatum. As PD progresses dopamine availability decreases, leading to the characteristic locomotor deficits including tremors, rigidity and bradykinesia (Chaudhuri et al., 2006). For several decades dopamine replacement therapy with l-dopa has been the gold-standard treatment for combating the motor symptoms for patients with PD. However, as disease progresses, l-dopa doses often need to be increased in order to manage symptoms. Approximately 52–78% of patients may in turn develop debilitating l-dopa induced dyskinesias (LIDs), classified as abnormalities or impairments of voluntary movements, within 10 years of initiating treatment (Manson et al., 2012). Accordingly, LIDs present a clinical-therapeutic conundrum, as the appearance of LIDs prevents further increasing l-dopa doses and in fact often needs to be managed by lowering l-dopa doses, which in turn leads to the loss of l-dopa’s anti-parkinsonian efficacy (Pandey and Srivanitchapoom, 2017).

To date, the only FDA approved therapy to combat LIDs in PD patients is amantadine. The clinical use of amantadine is unfortunately limited by several side effects, the development of tolerance and a lack of efficacy in some patients (Perez-Lloret and Rascol, 2018Sharma et al., 2018). For this reason, there is a great unmet clinical need for new therapies to treat LIDs.

In order to develop new therapies for LIDs, it is necessary to target the underlying mechanisms. Amantadine has been thought to exert its beneficial effects through its weak NMDA receptor antagonism at synapses (Blanpied et al., 2005Paquette et al., 2012), while recent research has intriguingly identified amantadine’s effects on glia as a potential mechanism (Kim et al., 2012Ossola et al., 2011). More generally, while there is no consensus, synapse loss and pathological regrowth (Fieblinger et al., 2014Suárez et al., 2014Zhang et al., 2013), changes in synaptic plasticity (Picconi et al., 2003Thiele et al., 2014) and neuroinflammation (Mulas et al., 2016), have all been implicated in LID pathogenesis. Given the growing understanding of the enumerate roles of glia in the healthy and diseased brain (Hammond et al., 2018Khakh and Sofroniew, 2015Morris et al., 2013) including LIDs (Mulas et al., 2016), targeting neuroinflammation, or perhaps more specifically glial signalling, provides a potential strategy for preclinical and clinical drug development for LIDs.

If targeting neuroinflammation, and or glial signalling, offers a potential strategy, then cannabinoid based therapies could be an option for treating LIDs. Cannabinoid-based therapies can exert effects on glia, are thought to suppress neuroinflammation, and have neuroprotective effects in preclinical animal models of several neurodegenerative disorders (Bisogno and Di Marzo, 2010). Intriguingly, some observational studies have indicated that smoking medical cannabis can alleviate LID in PD patients (Finseth et al., 2015Lotan et al., 2014Venderová et al., 2004). Cannabinoid effects are primarily mediated through the cannabinoid receptors CB1 and CB2 and previous preclinical studies have demonstrated that CB1 agonists (dos-Santos-Pereira et al., 2016Ferrer et al., 2003Fox et al., 2002Martinez et al., 2012Morgese et al., 2007Song et al., 2014Walsh et al., 2010) exert anti-dyskinetic properties. In contrast, the therapeutic potential of exclusively targeting CB2 receptors has not yet been investigated.

While CB2 selective agonists have not been investigated, they could be of particular clinical relevance, as it is suggested that targeting this receptor does not provoke the psychoactive side-effects associated with CB1 receptor activation (Pacher et al., 2006). Moreover, in the brain CB2 receptors are thought to be predominantly expressed by microglia (Jordan and Xi, 2019Palazuelos et al., 2009) and astrocytes (Fernández-Trapero et al., 2017). Further, while CB2 expression in the healthy brain is relatively low, expression in glia is elevated in preclinical animal models of neurodegenerative diseases as well as in human brain tissue of Parkinson’s (Gómez-Gálvez et al., 2016), Huntington’s (Palazuelos et al., 2009) and Alzheimer’s disease patients (Benito et al., 2003). One intriguing interpretation of this is that CB2 expression is upregulated as part of a glial homeostatic response. In support of this, CB2 receptor activation appears to initiate a signalling cascade in glia leading to decreased pro-inflammatory cytokine production and decreased glial cell-proliferation (Ashton and Glass, 2007). These effects are hypothesized to contribute to neuroprotection in various toxin based rodent models of PD including rotenone (Javed et al., 2016), MPTP (Price et al., 2009) and LPS (Gómez-Gálvez et al., 2016). Thus, pharmacologically stimulating CB2 receptor signalling may be a promising therapeutic strategy for neurodegenerative conditions where neuroinflammation, and therefore altered glial responses (Ben Haim et al., 2015Booth et al., 2017Morris et al., 2013Priller and Prinz, 2019), are implicated.

Collectively, the apparent potential of cannabinoid therapies for treatment of several neurological conditions (Benito et al., 2003Gómez-Gálvez et al., 2016Palazuelos et al., 2009), the putative relationship of neuroinflammation in LID pathogenesis (Mulas et al., 2016), the expression of CB2 in glia and their stated anti-inflammatory properties (Gómez-Gálvez et al., 2016Javed et al., 2016Price et al., 2009), all point to CB2 receptors as a promising therapeutic target for dyskinesia. Thus, in the current study, we hypothesized that a CB2 receptor agonist may exert anti-dyskinetic efficacy in a mouse model of LID.

To test this hypothesis, we utilized the selective CB2 receptor agonist HU-308 (Hanus et al., 1999). HU-308 treatment has previously been shown to reduce microglia proliferation and cytokine expression and provide concurrent neuroprotection in mouse models of Parkinson’s (Gómez-Gálvez et al., 2016) and Huntington’s disease (Palazuelos et al., 2009). We aimed to determine if the putative actions of HU-308 on glia could also translate into an effect on LIDs created by repeat l-dopa treatment in a 6-OHDA mouse model of PD. We also investigated the potential anti-dyskinetic effect of HU-308 alone and in combination with amantadine, as well as their effects on glial reactivity in striatal tissue of 6-OHDA lesioned mice expressing LIDs.


Throughout this study we have referred to the term neuroinflammation, which is classically defined by changes in glial proliferation, morphology and cytokine release, among other measures. However there has been an increasing recognition of the limitations of the term neuroinflammation and an increasing understanding of the multiple roles of glia in the healthy and diseased brain (Hammond et al., 2018Khakh and Sofroniew, 2015Morris et al., 2013). With this understanding, we and others have suggested that targeting glial homeostasis offers a promising route for treating neurodegenerative diseases and conditions in which synapse and neuron loss is implicated (Morris et al., 2013).

Given the evidence of altered glial function and morphology associated with LIDs (Mulas et al., 2016), the presence of CB2 receptors on glia, and their apparent effect of reversing altered glial function and morphology, (Benito et al., 2008), we hypothesized targeting CB2 receptors could provide a potential avenue for attenuating dyskinesia. Using a mouse model of LIDs, the current study revealed three key findings.

First, the CB2 selective agonist HU-308 dose-dependently reduced LID to the same magnitude as the current frontline treatment, amantadine. Second, treatment with HU-308 plus amantadine resulted in an additive anti-dyskinetic effect. Third, these treatment regimens decreased the expression of neuroinflammatory mediators in the striatum of 6-OHDA lesioned mice. Our findings therefore provide the first evidence that targeting CB2 receptors may be a promising pharmacological strategy for alleviating LIDs, a major unmet clinical need for PD patients.

HU-308 dose-dependently reduced AIMs

To favor drug safety and tolerability, and to avoid adverse effects, cannabinoid treatments are preferably administered at the lowest therapeutically efficacious dose (MacCallum and Russo, 2018). Thus, we first determined the dose-dependent effect of HU-308 on reducing AIMs in a mouse model of LID. Our results suggest that 2.5 mg/kg and 5 mg/kg of HU-308 were able to reduce dyskinesia’s to a similar extent, which is greater than that seen with 1 mg/kg HU-308. Collectively, these results allow us to conclude that 2.5 mg/kg HU-308 is an efficacious dose that achieves maximum reduction of AIMs in mice.

The anti-dyskinetic effect of HU-308 was CB2 specific

One major caveat for the therapeutic development of cannabinoids are the unwanted psychoactive side-effects associated with CB1 agonism (Pacher et al., 2006). A CB2 agonist offers a desirable alternative as it does not appear to trigger these side-effects (Tabrizi et al., 2016). HU-308 has previously been shown to be a CB2 specific agonist, efficiently binding to CB2 (Ki = 22.7 ± 3.9nM), while not binding to CB1 (Ki > 10 μM) (Hanus et al., 1999). In order to test receptor specificity of drugs, it is common practice to demonstrate a lack of effect in receptor knockout animals.

Although there are multiple CB2 knockout mouse lines available (Buckley et al., 2000Li and Kim, 2016), CB2 lacking mice may be more susceptible to toxins, as evidenced by an increase in lesion severity in the LPS mouse model of PD (Gómez-Gálvez et al., 2016). Thus, the lesion size following 6-OHDA treatment would likely be larger in CB2 knockouts compared to wildtype controls which would affect the LID magnitude and make results difficult to interpret.

Accordingly, as a consistent lesion volume is critical for our LID studies we were unable to use these genetically modified mice in our study. Therefore, in the current study we tested receptor specificity by determining if the selective CB2 receptor antagonist SR144528 can block the anti-dyskinetic effect of the CB2 receptor agonist HU-308 administered only after the lesion is created. This strategy has previously been used in a rat model of Huntington’s disease, with SR144528 blocking the neuroprotective effect of HU-308 (Sagredo et al., 2009).

As hypothesized, SR144528 by itself had no effect on AIMs, but when administered in conjunction with HU-308, SR144528 fully blocked the anti-dyskinetic effect of HU-308. Together, these results allow us to conclude that HU-308’s anti-dyskinetic properties are CB2 specific and unlikely due to any off-target effects.

The anti-dyskinetic effect of HU-308 is comparable to that of amantadine

After establishing the anti-dyskinetic efficacy of HU-308 we next aimed to compare the magnitude of this effect to that of amantadine. Others have previously reported that a dosage of 40 mg/kg amantadine is close to the upper limit of its therapeutic efficacy in rodents (Brigham et al., 2018Danysz et al., 1997).

In our hands, amantadine at this dose resulted in a 30% reduction of AIMs and 45% reduction in FosB expression, which closely aligns with previous studies in 6-OHDA lesioned mice reporting the ability of amantadine to reduce both dyskinetic behavior (up to 36%) (Sebastianutto et al., 2016) and FosB expression (up to 47%) (Doo et al., 2014). This demonstrated the robustness of this model as a tool to detect improvements in LID symptoms. Remarkably, 2.5 mg/kg HU-308 was as effective as amantadine and reduced AIMs by 31% and FosB expression by 50%.

If it were to reproduce in human cohorts, pharmacologically targeting CB2 might provide a useful alternative to amantadine, for example in cases where there are amantadine specific side effects. In support of this, several CB2 selective agonists have shown to be safe, well tolerated, not associated with any major side effects, and effective in treating peripheral pain and inflammatory conditions in Phase 1 and 2 clinical trials (Di Marzo, 2018Tabrizi et al., 2016). Accordingly, clinical trials investigating their efficacy for neurodegenerative diseases is currently in high demand.

HU-308 and amantadine have an additive anti-dyskinetic effect

To maximize symptomatic relief it is common practice to treat patients with a combination of different drugs (Thorlund and Mills, 2012). This is particularly valuable where two drugs can act more effectively together such that the effect of the two drugs in combination can exceed the maximal effect of either drug used alone. We demonstrated that the anti-dyskinetic effect of HU-308 is dose-dependent, but maximal at 2.5 mg/kg.

It is striking, therefore, that we found the addition of amantadine to HU-308 treatment resulted in a greater magnitude of reduction in AIM scores compared with that maximally achieved with HU-308. This result could be taken to suggest that HU-308 and amantadine ultimately each modulate the expression of LIDs through different but synergistic pathways so that the combined effect of both exceeds that which can be achieved by HU-308 alone. Regardless, our result suggests a combined HU-308 and amantadine treatment may be of greater benefit for PD patients with LIDs than either alone.

HU-308 and amantadine exert anti-inflammatory properties in striatal tissue

In investigating the apparent effects of amantadine and HU-308 on glial responses in this study we have largely confirmed previous published findings. In particular, it has been reported that amantadine has effects on glia independently of its actions on NMDA receptors, and that this is associated with the protection of cultured DA neurons against MPP+ and LPS toxicity (Kim et al., 2012Ossola et al., 2011) as well the protection of TH+ neurons in an MPTP and LPS mouse model (Kim et al., 2012).

The latter study demonstrated amantadine treatment reduced microglia proliferation and decreased NF-κB activity (Kim et al., 2012). Our data confirms and advances these findings. Whereas previous studies focused on the impact of amantadine on neuroinflammation and related degeneration of dopaminergic neurons in the SNpc, we are the first to report that amantadine has the capability to reduce microglial proliferation, GFAP+ astrocytes and cytokine release in the striatum of dyskinetic mice.

Our experiments also confirm the effect of HU-308 in striatal tissue, as previously demonstrated in Parkinson’s (Gómez-Gálvez et al., 2016) and Huntington’s (Palazuelos et al., 2009) disease mouse models. In particular, those studies demonstrated that HU-308 treated mice showed a reduction in microglial proliferaiton and GFAP+ astrocyte populations in an excitotoxic Huntington’s model (Palazuelos et al., 2009) and a reduction in activated microglia as well as reduced mRNA expression of the pro-inflammatory cytokines TNFα and IL-1β in a Parkinson’s model (Gómez-Gálvez et al., 2016). Accordingly, our results strengthen the hypothesis of an anti-inflammatory protential of HU-308 acrross multiple neurodegenerative disorders and models.

HU-308 and amantadine did not exhibit an additive effect on neuroinflammation

Despite finding that both HU-308 and amantadine exert significant effects on glia in our model, we did not find an additive effect of combined treatment. This finding is not surprising as we (Morris et al., 2013Morris et al., 2014) and others (Hammond et al., 2018Khakh and Sofroniew, 2015) have previously suggested that microglia and astrocytes are far more complex than previously thought. For example we now know that activated microglia drive the activation of astrocytes (Liddelow et al., 2017), that there are unique subsets of glia with only a proportion impacting neurodegenerative diseases (Deczkowska et al., 2018Jordão et al., 2019Keren-Shaul et al., 2017Masuda et al., 2019) and that immediate activation and proliferation of microglia after neuronal injury may favor recovery (Tay et al., 2017Tay et al., 2018). Accordingly, amantadine and HU-308 could be acting on some of these glial cell functions, and our broad measurements of glial cell counts and cytokine measurements only provide a small snapshot of the effects occurring in glia in our model.

Alternatively, the behavioural effect of each drug may be due, fully or in part, to effects of the drugs that are independent of their actions on dampening an inflammatory response associated with LIDs. For example, it has long been thought amantadine primarily suppresses LIDs via its weak NMDA antagonism (Blanpied et al., 2005Paquette et al., 2012) on striatal neurons, which may be separate, additional to, or part of, its reported anti-inflammatory actions. Meanwhile, the well known presence of CB2 receptors on glia does not rule out a potential direct or indirect action of CB2 receptor agonists at synapses.

Indeed, activation of CB2 receptors in the hippocampus for 7–10 days increases mEPSC frequency and spine density, suggesting CB2 receptors may also function to modulate synaptic activity (Kim and Li, 2015). The latter finding could result from an indirect action of CB2 receptor activation on glia, since recent research suggests that glia regulate synapses in healthy conditions and in disease (Morris et al., 2013).

Our measures in this paper are too rudimentary to explore these various mechanisms, and much further research is needed. However, not withstanding this limitation, our data adds weight to the concept that agents that act on glia may provide a promising option in preclinical and clinical drug development for neurodegenerative diseases generally and for LIDs in particular.

Strengths, limitations and future directions

The current study had several strengths, supporting the robustness of our findings. First, we ensured that anti-dyskinetic effects were not due to coincidental allocation of mice with a lesser lesion into any one particular treatment group, by confirming TH levels in striatal tissue were not different between groups. Second, by conducting our study in mice with established LIDs (as described previously by us (Rentsch et al., 2019) and others (Sebastianutto et al., 2016)), mice were distributed so that there was no coincidental allocation of mice with lower or higher average AIM score in one or another group prior to treatment.

Lastly, our behavioural data were largely correlated to the expression of FosB, a widely used molecular marker of LID (Andersson et al., 1999Lundblad et al., 2004Winkler et al., 2002). However, while this marker is mostly reliable in detecting gross changes in dyskinesia severity (dyskinetic vs. non-lesioned) it is possible this marker is unable to detect subtle to moderate changes within animals that are expressing dyskinesia (Smith et al., 2012). This may explain the instances in which FosB expression did not precisely corroborate our behavioural findings.

Our study also had limitations. As is often the case in preclinical research, our study was conducted in a homogenous population of adult male C57BL/6j mice. Thus, the potential efficacy of HU-308 has not yet been assessed in cohorts with different ages, sexes and strains. These are important next preclinical steps, before translation is considered. Furthermore, while cannabinoid treatments shape as promising therapeutic targets for motor disorders, an important caveat is that cannabinoids might also act as motor-depressants.

However, these effects are generally thought to be mediated by CB1 signalling, rather than CB2, which was one of the primary reasons we were interested in pursuing a CB2 agonist in this study (Hanus et al., 1999). In confirmation of this, in a previous study, HU-308 did not affect general motor activity in an open field test, nor did it cause catalepsy in naïve mice, even when administered at high doses (Hanus et al., 1999). Nevertheless, it will be important to conclusively determine if HU-308 has any effect on general or PD and LID specific motor activity in future studies. Finally, LIDs can last for many years in patients and we therefore suggest that, based on the enticing results of the current study, future studies of CB2 agonists in dyskinesias should confirm efficacy over a greatly extended period.


Collectively, our findings suggest CB2 agonists offer a putative target to treat LIDs, with efficacy comparable to the frontline treatment amantadine. This behavioural effect is associated with an effect on glial signalling (as evidenced by downregulation of neuroinflammation), providing further evidence that therapeutics targeting neuroinflammation and/or glial homeostasis may provide benefit for combating LIDs. Furthermore, one of the more important findings was the demonstration of an additive effect of HU-308 and amantadine that is greater than that achieved with HU-308 alone.

Although we do not yet know the precise mechanisms driving this effect, our results suggest they may act by different but synergistic actions which has important clinical implications. We have suggested several exciting future directions to investigate the mechanism by which amantadine and HU-308 may exert their effects, particularly exploring novel features of microglia and astrocyte physiology and pathophysiology and their direct and/or indirect impact on neuronal synaptic signalling which is known to be altered in dyskinesias. Our study suggests that targeting glial function may be an important strategy for developing therapies for treating LIDs, a major unmet need for PD patients.

University of Technology Sydney
Media Contacts:
Marea Martlew – University of Technology Sydney
Image Source:
The image is in the public domain.

Original Research: Open access
“Targeting the cannabinoid receptor CB2 in a mouse model of l-dopa induced Dyskinesia”. Bryce Vissel et al.
Neurobiology of Disease doi:10.1016/j.nbd.2019.104646.


Please enter your comment!
Please enter your name here

Questo sito usa Akismet per ridurre lo spam. Scopri come i tuoi dati vengono elaborati.